Google applies search technologies behind the firewall.
By Kevin Davies
April 1, 2008 | In a recent 60 Minutes broadcast about the cost of minting the penny, an MIT student named Jeff Glore calculated that Americans spend an average of 2.4 hours a year fumbling and searching for those annoying 1-cent coins.
Most biomedical researchers would be thrilled if that was all the time they spent on fruitless searches for data, protocols, clinical trial, legal, and regulatory documents. According Jim Golden, SAIC's CTO Life Sciences, one pharma insider estimates that researchers spend 15 minutes per person per day tracking information. Says Golden, that's the equivalent of 333 out of 10,000 employees doing nothing but search! The lost productivity and delays in drug development and regulatory submissions potentially tally tens of millions of dollars a year.
A new arena called enterprise search can potentially help researchers and executives instantly find documents within the enterprise through a simple search. Now the field had a formidable new entrant - Google. The search giant offers a search appliance that allows users to search all enterprise content - including intranets, databases, file shares, real-time business data, and content management systems - through one simple search box.
"The enterprise is very much a key component to Google's overall business," says Mark Rudick, Google's head of sales engineering for North America. "There is exponentially more [relevant] information within the corporate firewall" than on the web. Google Enterprise has more than 600 staff, and is the fastest growing segment within Google, with more than 12,000 customers.
Data Silos
Much of the problem is that life science organizations are collections by information repositories, or data silos, in which "valuable stuff goes in and never comes out again," says Golden. "They're not connected to each other; there's also no context and semantics to the information that's in them."
Moreover, when a pharma launches a drug, it typically creates a "defense database" of relevant legal, regulatory, experimental, and clinical data. "This process can cost a pharma company anywhere between $30-50 million per drug," says Golden. "If we actually had search and semantic interoperability... this process of creating the defense database would have started the first time we did an experiment looking at a target for a drug, not at the end, when we have to spend a lot of time going back and figuring out what we did."
The goal of Google's enterprise appliance is to have a search go across the organization. "We don't care where the information lives, we just need to be able to get to it," says Rudick. Google's application can dip into various data repositories including databases and content management systems such as Documentum and Sharepoint.
"The reason you buy a search engine is to get the right results at the end of the day," says Rudick. "We've leveraged the [more than 100] algorithms on google.com, [but] we've tuned them differently, and added new algorithms to be more applicable to the enterprise." In some cases, Rudick says it is possible to do some specific tuning or tweaking, based on data sources or dates, or adding key matches. But "95% of the time with the search appliance, that's not necessary," he says. A single Google appliance can scale up to 24 queries per second, which provides ample capacity.
Security is a key issue, ensuring users are granted appropriate access to content. Google leverages any security infrastructure the organization already has in place. "When the appliance serves up a result, it makes sure that it only serves up the results to the users that they are authorized to see in the first place," says Rudick.
Google has only recently begun targeting pharmaceutical companies, but it is already gaining traction. One global big pharma (headquartered in New Jersey) worked with Google in four major product areas, including data redundancy, product management, locating internal talent for recruitment, and marketing - managing multiple repositories, emails, databases, and hard drives, to help consolidate information.
"The overall impact that we saw was around 70% workers saw improved productivity in their day-to-day behavior by having a system that brought relevant information to them more quickly," says Amy Kim, sales manager for Google Enterprise. She adds that the appliance is equally well suited for smaller departmental purposes as well as enterprise search.
Kim says the entry level cost for the appliance within a firewall is $30,000 for a two-year agreement to index 500,000 documents.
___________________________________________________
This article appeared in Bio-IT World Magazine.
Subscriptions are free for qualifying individuals. Apply Today.